What happens when AI starts moving faster than the people meant to control it?

In this episode, I'm joined by Bernard Montel, Field CTO EMEA at Tenable, for a timely conversation about the AI risks many organizations may be underestimating. Bernard believes we are heading toward a defining AI accident and that the first major incident may come through speed, scale, and unintended consequences rather than a malicious attack.

We talk about why so many companies feel pressure to adopt AI at pace, while visibility, governance, and control struggle to keep up. Bernard describes this moment as "driving faster than we can steer," and explains why shadow AI, overprivileged identities, cloud misconfigurations, and exposed AI projects are already creating real business risk.

The conversation also looks at agentic AI and why giving systems the ability to take action changes the security equation. A chatbot giving a wrong answer is one problem. An AI agent making flawed decisions, leaking data, or interacting with industrial systems is something very different.

Bernard also shares why AI can become a distraction from the security basics that still matter, including cloud security, identity, exposure management, and vulnerability remediation. Attackers may be using AI to move faster, but many of the weaknesses they exploit remain painfully familiar.

We also discuss Tenable's new agentic AI framework, announced during RSA, and how the company is using AI to help security teams respond at machine speed while reducing exposure across IT, cloud, OT, identity, and AI environments.

For business and security leaders, this episode offers a clear warning and a practical takeaway. AI adoption is no longer a future conversation, but control, governance, and exposure management need to move with it.

How prepared is your organization for an AI incident caused by accident rather than attack? Share your thoughts.

Useful Links

Visit the Sponsors of Tech Talks Network and learn more about the NordLayer Browser.